Estimating risk‐return relationships: An analysis of measures